1use anyhow::Context;
2use globset::{Glob, GlobSet, GlobSetBuilder};
3use itertools::Itertools;
4use regex::Regex;
5use serde::{Deserialize, Serialize};
6use std::cmp::Ordering;
7use std::error::Error;
8use std::fmt::{Display, Formatter};
9use std::mem;
10use std::path::StripPrefixError;
11use std::sync::{Arc, OnceLock};
12use std::{
13 ffi::OsStr,
14 path::{Path, PathBuf},
15 sync::LazyLock,
16};
17
18use crate::{rel_path::RelPath, shell::ShellKind};
19
20static HOME_DIR: OnceLock<PathBuf> = OnceLock::new();
21
22/// Returns the path to the user's home directory.
23pub fn home_dir() -> &'static PathBuf {
24 HOME_DIR.get_or_init(|| {
25 if cfg!(any(test, feature = "test-support")) {
26 if cfg!(target_os = "macos") {
27 PathBuf::from("/Users/zed")
28 } else if cfg!(target_os = "windows") {
29 PathBuf::from("C:\\Users\\zed")
30 } else {
31 PathBuf::from("/home/zed")
32 }
33 } else {
34 dirs::home_dir().expect("failed to determine home directory")
35 }
36 })
37}
38
39pub trait PathExt {
40 /// Compacts a given file path by replacing the user's home directory
41 /// prefix with a tilde (`~`).
42 ///
43 /// # Returns
44 ///
45 /// * A `PathBuf` containing the compacted file path. If the input path
46 /// does not have the user's home directory prefix, or if we are not on
47 /// Linux or macOS, the original path is returned unchanged.
48 fn compact(&self) -> PathBuf;
49
50 /// Returns a file's extension or, if the file is hidden, its name without the leading dot
51 fn extension_or_hidden_file_name(&self) -> Option<&str>;
52
53 fn try_from_bytes<'a>(bytes: &'a [u8]) -> anyhow::Result<Self>
54 where
55 Self: From<&'a Path>,
56 {
57 #[cfg(unix)]
58 {
59 use std::os::unix::prelude::OsStrExt;
60 Ok(Self::from(Path::new(OsStr::from_bytes(bytes))))
61 }
62 #[cfg(windows)]
63 {
64 use tendril::fmt::{Format, WTF8};
65 WTF8::validate(bytes)
66 .then(|| {
67 // Safety: bytes are valid WTF-8 sequence.
68 Self::from(Path::new(unsafe {
69 OsStr::from_encoded_bytes_unchecked(bytes)
70 }))
71 })
72 .with_context(|| format!("Invalid WTF-8 sequence: {bytes:?}"))
73 }
74 }
75
76 /// Converts a local path to one that can be used inside of WSL.
77 /// Returns `None` if the path cannot be converted into a WSL one (network share).
78 fn local_to_wsl(&self) -> Option<PathBuf>;
79
80 /// Returns a file's "full" joined collection of extensions, in the case where a file does not
81 /// just have a singular extension but instead has multiple (e.g File.tar.gz, Component.stories.tsx)
82 ///
83 /// Will provide back the extensions joined together such as tar.gz or stories.tsx
84 fn multiple_extensions(&self) -> Option<String>;
85
86 /// Try to make a shell-safe representation of the path.
87 fn try_shell_safe(&self, shell_kind: ShellKind) -> anyhow::Result<String>;
88}
89
90impl<T: AsRef<Path>> PathExt for T {
91 fn compact(&self) -> PathBuf {
92 if cfg!(any(target_os = "linux", target_os = "freebsd")) || cfg!(target_os = "macos") {
93 match self.as_ref().strip_prefix(home_dir().as_path()) {
94 Ok(relative_path) => {
95 let mut shortened_path = PathBuf::new();
96 shortened_path.push("~");
97 shortened_path.push(relative_path);
98 shortened_path
99 }
100 Err(_) => self.as_ref().to_path_buf(),
101 }
102 } else {
103 self.as_ref().to_path_buf()
104 }
105 }
106
107 fn extension_or_hidden_file_name(&self) -> Option<&str> {
108 let path = self.as_ref();
109 let file_name = path.file_name()?.to_str()?;
110 if file_name.starts_with('.') {
111 return file_name.strip_prefix('.');
112 }
113
114 path.extension()
115 .and_then(|e| e.to_str())
116 .or_else(|| path.file_stem()?.to_str())
117 }
118
119 fn local_to_wsl(&self) -> Option<PathBuf> {
120 // quite sketchy to convert this back to path at the end, but a lot of functions only accept paths
121 // todo: ideally rework them..?
122 let mut new_path = std::ffi::OsString::new();
123 for component in self.as_ref().components() {
124 match component {
125 std::path::Component::Prefix(prefix) => {
126 let drive_letter = prefix.as_os_str().to_string_lossy().to_lowercase();
127 let drive_letter = drive_letter.strip_suffix(':')?;
128
129 new_path.push(format!("/mnt/{}", drive_letter));
130 }
131 std::path::Component::RootDir => {}
132 std::path::Component::CurDir => {
133 new_path.push("/.");
134 }
135 std::path::Component::ParentDir => {
136 new_path.push("/..");
137 }
138 std::path::Component::Normal(os_str) => {
139 new_path.push("/");
140 new_path.push(os_str);
141 }
142 }
143 }
144
145 Some(new_path.into())
146 }
147
148 fn multiple_extensions(&self) -> Option<String> {
149 let path = self.as_ref();
150 let file_name = path.file_name()?.to_str()?;
151
152 let parts: Vec<&str> = file_name
153 .split('.')
154 // Skip the part with the file name extension
155 .skip(1)
156 .collect();
157
158 if parts.len() < 2 {
159 return None;
160 }
161
162 Some(parts.into_iter().join("."))
163 }
164
165 fn try_shell_safe(&self, shell_kind: ShellKind) -> anyhow::Result<String> {
166 let path_str = self
167 .as_ref()
168 .to_str()
169 .with_context(|| "Path contains invalid UTF-8")?;
170 shell_kind
171 .try_quote(path_str)
172 .as_deref()
173 .map(ToOwned::to_owned)
174 .context("Failed to quote path")
175 }
176}
177
178pub fn path_ends_with(base: &Path, suffix: &Path) -> bool {
179 strip_path_suffix(base, suffix).is_some()
180}
181
182pub fn strip_path_suffix<'a>(base: &'a Path, suffix: &Path) -> Option<&'a Path> {
183 if let Some(remainder) = base
184 .as_os_str()
185 .as_encoded_bytes()
186 .strip_suffix(suffix.as_os_str().as_encoded_bytes())
187 {
188 if remainder
189 .last()
190 .is_none_or(|last_byte| std::path::is_separator(*last_byte as char))
191 {
192 let os_str = unsafe {
193 OsStr::from_encoded_bytes_unchecked(
194 &remainder[0..remainder.len().saturating_sub(1)],
195 )
196 };
197 return Some(Path::new(os_str));
198 }
199 }
200 None
201}
202
203/// In memory, this is identical to `Path`. On non-Windows conversions to this type are no-ops. On
204/// windows, these conversions sanitize UNC paths by removing the `\\\\?\\` prefix.
205#[derive(Eq, PartialEq, Hash, Ord, PartialOrd)]
206#[repr(transparent)]
207pub struct SanitizedPath(Path);
208
209impl SanitizedPath {
210 pub fn new<T: AsRef<Path> + ?Sized>(path: &T) -> &Self {
211 #[cfg(not(target_os = "windows"))]
212 return Self::unchecked_new(path.as_ref());
213
214 #[cfg(target_os = "windows")]
215 return Self::unchecked_new(dunce::simplified(path.as_ref()));
216 }
217
218 pub fn unchecked_new<T: AsRef<Path> + ?Sized>(path: &T) -> &Self {
219 // safe because `Path` and `SanitizedPath` have the same repr and Drop impl
220 unsafe { mem::transmute::<&Path, &Self>(path.as_ref()) }
221 }
222
223 pub fn from_arc(path: Arc<Path>) -> Arc<Self> {
224 // safe because `Path` and `SanitizedPath` have the same repr and Drop impl
225 #[cfg(not(target_os = "windows"))]
226 return unsafe { mem::transmute::<Arc<Path>, Arc<Self>>(path) };
227
228 // TODO: could avoid allocating here if dunce::simplified results in the same path
229 #[cfg(target_os = "windows")]
230 return Self::new(&path).into();
231 }
232
233 pub fn new_arc<T: AsRef<Path> + ?Sized>(path: &T) -> Arc<Self> {
234 Self::new(path).into()
235 }
236
237 pub fn cast_arc(path: Arc<Self>) -> Arc<Path> {
238 // safe because `Path` and `SanitizedPath` have the same repr and Drop impl
239 unsafe { mem::transmute::<Arc<Self>, Arc<Path>>(path) }
240 }
241
242 pub fn cast_arc_ref(path: &Arc<Self>) -> &Arc<Path> {
243 // safe because `Path` and `SanitizedPath` have the same repr and Drop impl
244 unsafe { mem::transmute::<&Arc<Self>, &Arc<Path>>(path) }
245 }
246
247 pub fn starts_with(&self, prefix: &Self) -> bool {
248 self.0.starts_with(&prefix.0)
249 }
250
251 pub fn as_path(&self) -> &Path {
252 &self.0
253 }
254
255 pub fn file_name(&self) -> Option<&std::ffi::OsStr> {
256 self.0.file_name()
257 }
258
259 pub fn extension(&self) -> Option<&std::ffi::OsStr> {
260 self.0.extension()
261 }
262
263 pub fn join<P: AsRef<Path>>(&self, path: P) -> PathBuf {
264 self.0.join(path)
265 }
266
267 pub fn parent(&self) -> Option<&Self> {
268 self.0.parent().map(Self::unchecked_new)
269 }
270
271 pub fn strip_prefix(&self, base: &Self) -> Result<&Path, StripPrefixError> {
272 self.0.strip_prefix(base.as_path())
273 }
274
275 pub fn to_str(&self) -> Option<&str> {
276 self.0.to_str()
277 }
278
279 pub fn to_path_buf(&self) -> PathBuf {
280 self.0.to_path_buf()
281 }
282}
283
284impl std::fmt::Debug for SanitizedPath {
285 fn fmt(&self, formatter: &mut Formatter<'_>) -> std::fmt::Result {
286 std::fmt::Debug::fmt(&self.0, formatter)
287 }
288}
289
290impl Display for SanitizedPath {
291 fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
292 write!(f, "{}", self.0.display())
293 }
294}
295
296impl From<&SanitizedPath> for Arc<SanitizedPath> {
297 fn from(sanitized_path: &SanitizedPath) -> Self {
298 let path: Arc<Path> = sanitized_path.0.into();
299 // safe because `Path` and `SanitizedPath` have the same repr and Drop impl
300 unsafe { mem::transmute(path) }
301 }
302}
303
304impl From<&SanitizedPath> for PathBuf {
305 fn from(sanitized_path: &SanitizedPath) -> Self {
306 sanitized_path.as_path().into()
307 }
308}
309
310impl AsRef<Path> for SanitizedPath {
311 fn as_ref(&self) -> &Path {
312 &self.0
313 }
314}
315
316#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
317pub enum PathStyle {
318 Posix,
319 Windows,
320}
321
322impl PathStyle {
323 #[cfg(target_os = "windows")]
324 pub const fn local() -> Self {
325 PathStyle::Windows
326 }
327
328 #[cfg(not(target_os = "windows"))]
329 pub const fn local() -> Self {
330 PathStyle::Posix
331 }
332
333 #[inline]
334 pub fn separator(&self) -> &'static str {
335 match self {
336 PathStyle::Posix => "/",
337 PathStyle::Windows => "\\",
338 }
339 }
340
341 pub fn is_windows(&self) -> bool {
342 *self == PathStyle::Windows
343 }
344
345 pub fn join(self, left: impl AsRef<Path>, right: impl AsRef<Path>) -> Option<String> {
346 let right = right.as_ref().to_str()?;
347 if is_absolute(right, self) {
348 return None;
349 }
350 let left = left.as_ref().to_str()?;
351 if left.is_empty() {
352 Some(right.into())
353 } else {
354 Some(format!(
355 "{left}{}{right}",
356 if left.ends_with(self.separator()) {
357 ""
358 } else {
359 self.separator()
360 }
361 ))
362 }
363 }
364
365 pub fn split(self, path_like: &str) -> (Option<&str>, &str) {
366 let Some(pos) = path_like.rfind(self.separator()) else {
367 return (None, path_like);
368 };
369 let filename_start = pos + self.separator().len();
370 (
371 Some(&path_like[..filename_start]),
372 &path_like[filename_start..],
373 )
374 }
375}
376
377#[derive(Debug, Clone)]
378pub struct RemotePathBuf {
379 style: PathStyle,
380 string: String,
381}
382
383impl RemotePathBuf {
384 pub fn new(string: String, style: PathStyle) -> Self {
385 Self { style, string }
386 }
387
388 pub fn from_str(path: &str, style: PathStyle) -> Self {
389 Self::new(path.to_string(), style)
390 }
391
392 pub fn path_style(&self) -> PathStyle {
393 self.style
394 }
395
396 pub fn to_proto(self) -> String {
397 self.string
398 }
399}
400
401impl Display for RemotePathBuf {
402 fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
403 write!(f, "{}", self.string)
404 }
405}
406
407pub fn is_absolute(path_like: &str, path_style: PathStyle) -> bool {
408 path_like.starts_with('/')
409 || path_style == PathStyle::Windows
410 && (path_like.starts_with('\\')
411 || path_like
412 .chars()
413 .next()
414 .is_some_and(|c| c.is_ascii_alphabetic())
415 && path_like[1..]
416 .strip_prefix(':')
417 .is_some_and(|path| path.starts_with('/') || path.starts_with('\\')))
418}
419
420#[derive(Debug, PartialEq)]
421#[non_exhaustive]
422pub struct NormalizeError;
423
424impl Error for NormalizeError {}
425
426impl std::fmt::Display for NormalizeError {
427 fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
428 f.write_str("parent reference `..` points outside of base directory")
429 }
430}
431
432/// Copied from stdlib where it's unstable.
433///
434/// Normalize a path, including `..` without traversing the filesystem.
435///
436/// Returns an error if normalization would leave leading `..` components.
437///
438/// <div class="warning">
439///
440/// This function always resolves `..` to the "lexical" parent.
441/// That is "a/b/../c" will always resolve to `a/c` which can change the meaning of the path.
442/// In particular, `a/c` and `a/b/../c` are distinct on many systems because `b` may be a symbolic link, so its parent isn't `a`.
443///
444/// </div>
445///
446/// [`path::absolute`](absolute) is an alternative that preserves `..`.
447/// Or [`Path::canonicalize`] can be used to resolve any `..` by querying the filesystem.
448pub fn normalize_lexically(path: &Path) -> Result<PathBuf, NormalizeError> {
449 use std::path::Component;
450
451 let mut lexical = PathBuf::new();
452 let mut iter = path.components().peekable();
453
454 // Find the root, if any, and add it to the lexical path.
455 // Here we treat the Windows path "C:\" as a single "root" even though
456 // `components` splits it into two: (Prefix, RootDir).
457 let root = match iter.peek() {
458 Some(Component::ParentDir) => return Err(NormalizeError),
459 Some(p @ Component::RootDir) | Some(p @ Component::CurDir) => {
460 lexical.push(p);
461 iter.next();
462 lexical.as_os_str().len()
463 }
464 Some(Component::Prefix(prefix)) => {
465 lexical.push(prefix.as_os_str());
466 iter.next();
467 if let Some(p @ Component::RootDir) = iter.peek() {
468 lexical.push(p);
469 iter.next();
470 }
471 lexical.as_os_str().len()
472 }
473 None => return Ok(PathBuf::new()),
474 Some(Component::Normal(_)) => 0,
475 };
476
477 for component in iter {
478 match component {
479 Component::RootDir => unreachable!(),
480 Component::Prefix(_) => return Err(NormalizeError),
481 Component::CurDir => continue,
482 Component::ParentDir => {
483 // It's an error if ParentDir causes us to go above the "root".
484 if lexical.as_os_str().len() == root {
485 return Err(NormalizeError);
486 } else {
487 lexical.pop();
488 }
489 }
490 Component::Normal(path) => lexical.push(path),
491 }
492 }
493 Ok(lexical)
494}
495
496/// A delimiter to use in `path_query:row_number:column_number` strings parsing.
497pub const FILE_ROW_COLUMN_DELIMITER: char = ':';
498
499const ROW_COL_CAPTURE_REGEX: &str = r"(?xs)
500 ([^\(]+)\:(?:
501 \((\d+)[,:](\d+)\) # filename:(row,column), filename:(row:column)
502 |
503 \((\d+)\)() # filename:(row)
504 )
505 |
506 ([^\(]+)(?:
507 \((\d+)[,:](\d+)\) # filename(row,column), filename(row:column)
508 |
509 \((\d+)\)() # filename(row)
510 )
511 |
512 (.+?)(?:
513 \:+(\d+)\:(\d+)\:*$ # filename:row:column
514 |
515 \:+(\d+)\:*()$ # filename:row
516 |
517 \:+()()$
518 )";
519
520/// A representation of a path-like string with optional row and column numbers.
521/// Matching values example: `te`, `test.rs:22`, `te:22:5`, `test.c(22)`, `test.c(22,5)`etc.
522#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Hash)]
523pub struct PathWithPosition {
524 pub path: PathBuf,
525 pub row: Option<u32>,
526 // Absent if row is absent.
527 pub column: Option<u32>,
528}
529
530impl PathWithPosition {
531 /// Returns a PathWithPosition from a path.
532 pub fn from_path(path: PathBuf) -> Self {
533 Self {
534 path,
535 row: None,
536 column: None,
537 }
538 }
539
540 /// Parses a string that possibly has `:row:column` or `(row, column)` suffix.
541 /// Parenthesis format is used by [MSBuild](https://learn.microsoft.com/en-us/visualstudio/msbuild/msbuild-diagnostic-format-for-tasks) compatible tools
542 /// Ignores trailing `:`s, so `test.rs:22:` is parsed as `test.rs:22`.
543 /// If the suffix parsing fails, the whole string is parsed as a path.
544 ///
545 /// Be mindful that `test_file:10:1:` is a valid posix filename.
546 /// `PathWithPosition` class assumes that the ending position-like suffix is **not** part of the filename.
547 ///
548 /// # Examples
549 ///
550 /// ```
551 /// # use util::paths::PathWithPosition;
552 /// # use std::path::PathBuf;
553 /// assert_eq!(PathWithPosition::parse_str("test_file"), PathWithPosition {
554 /// path: PathBuf::from("test_file"),
555 /// row: None,
556 /// column: None,
557 /// });
558 /// assert_eq!(PathWithPosition::parse_str("test_file:10"), PathWithPosition {
559 /// path: PathBuf::from("test_file"),
560 /// row: Some(10),
561 /// column: None,
562 /// });
563 /// assert_eq!(PathWithPosition::parse_str("test_file.rs"), PathWithPosition {
564 /// path: PathBuf::from("test_file.rs"),
565 /// row: None,
566 /// column: None,
567 /// });
568 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:1"), PathWithPosition {
569 /// path: PathBuf::from("test_file.rs"),
570 /// row: Some(1),
571 /// column: None,
572 /// });
573 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:1:2"), PathWithPosition {
574 /// path: PathBuf::from("test_file.rs"),
575 /// row: Some(1),
576 /// column: Some(2),
577 /// });
578 /// ```
579 ///
580 /// # Expected parsing results when encounter ill-formatted inputs.
581 /// ```
582 /// # use util::paths::PathWithPosition;
583 /// # use std::path::PathBuf;
584 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:a"), PathWithPosition {
585 /// path: PathBuf::from("test_file.rs:a"),
586 /// row: None,
587 /// column: None,
588 /// });
589 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:a:b"), PathWithPosition {
590 /// path: PathBuf::from("test_file.rs:a:b"),
591 /// row: None,
592 /// column: None,
593 /// });
594 /// assert_eq!(PathWithPosition::parse_str("test_file.rs"), PathWithPosition {
595 /// path: PathBuf::from("test_file.rs"),
596 /// row: None,
597 /// column: None,
598 /// });
599 /// assert_eq!(PathWithPosition::parse_str("test_file.rs::1"), PathWithPosition {
600 /// path: PathBuf::from("test_file.rs"),
601 /// row: Some(1),
602 /// column: None,
603 /// });
604 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:1::"), PathWithPosition {
605 /// path: PathBuf::from("test_file.rs"),
606 /// row: Some(1),
607 /// column: None,
608 /// });
609 /// assert_eq!(PathWithPosition::parse_str("test_file.rs::1:2"), PathWithPosition {
610 /// path: PathBuf::from("test_file.rs"),
611 /// row: Some(1),
612 /// column: Some(2),
613 /// });
614 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:1::2"), PathWithPosition {
615 /// path: PathBuf::from("test_file.rs:1"),
616 /// row: Some(2),
617 /// column: None,
618 /// });
619 /// assert_eq!(PathWithPosition::parse_str("test_file.rs:1:2:3"), PathWithPosition {
620 /// path: PathBuf::from("test_file.rs:1"),
621 /// row: Some(2),
622 /// column: Some(3),
623 /// });
624 /// ```
625 pub fn parse_str(s: &str) -> Self {
626 let trimmed = s.trim();
627 let path = Path::new(trimmed);
628 let maybe_file_name_with_row_col = path.file_name().unwrap_or_default().to_string_lossy();
629 if maybe_file_name_with_row_col.is_empty() {
630 return Self {
631 path: Path::new(s).to_path_buf(),
632 row: None,
633 column: None,
634 };
635 }
636
637 // Let's avoid repeated init cost on this. It is subject to thread contention, but
638 // so far this code isn't called from multiple hot paths. Getting contention here
639 // in the future seems unlikely.
640 static SUFFIX_RE: LazyLock<Regex> =
641 LazyLock::new(|| Regex::new(ROW_COL_CAPTURE_REGEX).unwrap());
642 match SUFFIX_RE
643 .captures(&maybe_file_name_with_row_col)
644 .map(|caps| caps.extract())
645 {
646 Some((_, [file_name, maybe_row, maybe_column])) => {
647 let row = maybe_row.parse::<u32>().ok();
648 let column = maybe_column.parse::<u32>().ok();
649
650 let suffix_length = maybe_file_name_with_row_col.len() - file_name.len();
651 let path_without_suffix = &trimmed[..trimmed.len() - suffix_length];
652
653 Self {
654 path: Path::new(path_without_suffix).to_path_buf(),
655 row,
656 column,
657 }
658 }
659 None => {
660 // The `ROW_COL_CAPTURE_REGEX` deals with separated digits only,
661 // but in reality there could be `foo/bar.py:22:in` inputs which we want to match too.
662 // The regex mentioned is not very extendable with "digit or random string" checks, so do this here instead.
663 let delimiter = ':';
664 let mut path_parts = s
665 .rsplitn(3, delimiter)
666 .collect::<Vec<_>>()
667 .into_iter()
668 .rev()
669 .fuse();
670 let mut path_string = path_parts.next().expect("rsplitn should have the rest of the string as its last parameter that we reversed").to_owned();
671 let mut row = None;
672 let mut column = None;
673 if let Some(maybe_row) = path_parts.next() {
674 if let Ok(parsed_row) = maybe_row.parse::<u32>() {
675 row = Some(parsed_row);
676 if let Some(parsed_column) = path_parts
677 .next()
678 .and_then(|maybe_col| maybe_col.parse::<u32>().ok())
679 {
680 column = Some(parsed_column);
681 }
682 } else {
683 path_string.push(delimiter);
684 path_string.push_str(maybe_row);
685 }
686 }
687 for split in path_parts {
688 path_string.push(delimiter);
689 path_string.push_str(split);
690 }
691
692 Self {
693 path: PathBuf::from(path_string),
694 row,
695 column,
696 }
697 }
698 }
699 }
700
701 pub fn map_path<E>(
702 self,
703 mapping: impl FnOnce(PathBuf) -> Result<PathBuf, E>,
704 ) -> Result<PathWithPosition, E> {
705 Ok(PathWithPosition {
706 path: mapping(self.path)?,
707 row: self.row,
708 column: self.column,
709 })
710 }
711
712 pub fn to_string(&self, path_to_string: impl Fn(&PathBuf) -> String) -> String {
713 let path_string = path_to_string(&self.path);
714 if let Some(row) = self.row {
715 if let Some(column) = self.column {
716 format!("{path_string}:{row}:{column}")
717 } else {
718 format!("{path_string}:{row}")
719 }
720 } else {
721 path_string
722 }
723 }
724}
725
726#[derive(Clone, Debug)]
727pub struct PathMatcher {
728 sources: Vec<String>,
729 glob: GlobSet,
730 path_style: PathStyle,
731}
732
733// impl std::fmt::Display for PathMatcher {
734// fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
735// self.sources.fmt(f)
736// }
737// }
738
739impl PartialEq for PathMatcher {
740 fn eq(&self, other: &Self) -> bool {
741 self.sources.eq(&other.sources)
742 }
743}
744
745impl Eq for PathMatcher {}
746
747impl PathMatcher {
748 pub fn new(
749 globs: impl IntoIterator<Item = impl AsRef<str>>,
750 path_style: PathStyle,
751 ) -> Result<Self, globset::Error> {
752 let globs = globs
753 .into_iter()
754 .map(|as_str| Glob::new(as_str.as_ref()))
755 .collect::<Result<Vec<_>, _>>()?;
756 let sources = globs.iter().map(|glob| glob.glob().to_owned()).collect();
757 let mut glob_builder = GlobSetBuilder::new();
758 for single_glob in globs {
759 glob_builder.add(single_glob);
760 }
761 let glob = glob_builder.build()?;
762 Ok(PathMatcher {
763 glob,
764 sources,
765 path_style,
766 })
767 }
768
769 pub fn sources(&self) -> &[String] {
770 &self.sources
771 }
772
773 pub fn is_match<P: AsRef<Path>>(&self, other: P) -> bool {
774 let other_path = other.as_ref();
775 self.sources.iter().any(|source| {
776 let as_bytes = other_path.as_os_str().as_encoded_bytes();
777 as_bytes.starts_with(source.as_bytes()) || as_bytes.ends_with(source.as_bytes())
778 }) || self.glob.is_match(other_path)
779 || self.check_with_end_separator(other_path)
780 }
781
782 fn check_with_end_separator(&self, path: &Path) -> bool {
783 let path_str = path.to_string_lossy();
784 let separator = self.path_style.separator();
785 if path_str.ends_with(separator) {
786 false
787 } else {
788 self.glob.is_match(path_str.to_string() + separator)
789 }
790 }
791}
792
793impl Default for PathMatcher {
794 fn default() -> Self {
795 Self {
796 path_style: PathStyle::local(),
797 glob: GlobSet::empty(),
798 sources: vec![],
799 }
800 }
801}
802
803/// Compares two sequences of consecutive digits for natural sorting.
804///
805/// This function is a core component of natural sorting that handles numeric comparison
806/// in a way that feels natural to humans. It extracts and compares consecutive digit
807/// sequences from two iterators, handling various cases like leading zeros and very large numbers.
808///
809/// # Behavior
810///
811/// The function implements the following comparison rules:
812/// 1. Different numeric values: Compares by actual numeric value (e.g., "2" < "10")
813/// 2. Leading zeros: When values are equal, longer sequence wins (e.g., "002" > "2")
814/// 3. Large numbers: Falls back to string comparison for numbers that would overflow u128
815///
816/// # Examples
817///
818/// ```text
819/// "1" vs "2" -> Less (different values)
820/// "2" vs "10" -> Less (numeric comparison)
821/// "002" vs "2" -> Greater (leading zeros)
822/// "10" vs "010" -> Less (leading zeros)
823/// "999..." vs "1000..." -> Less (large number comparison)
824/// ```
825///
826/// # Implementation Details
827///
828/// 1. Extracts consecutive digits into strings
829/// 2. Compares sequence lengths for leading zero handling
830/// 3. For equal lengths, compares digit by digit
831/// 4. For different lengths:
832/// - Attempts numeric comparison first (for numbers up to 2^128 - 1)
833/// - Falls back to string comparison if numbers would overflow
834///
835/// The function advances both iterators past their respective numeric sequences,
836/// regardless of the comparison result.
837fn compare_numeric_segments<I>(
838 a_iter: &mut std::iter::Peekable<I>,
839 b_iter: &mut std::iter::Peekable<I>,
840) -> Ordering
841where
842 I: Iterator<Item = char>,
843{
844 // Collect all consecutive digits into strings
845 let mut a_num_str = String::new();
846 let mut b_num_str = String::new();
847
848 while let Some(&c) = a_iter.peek() {
849 if !c.is_ascii_digit() {
850 break;
851 }
852
853 a_num_str.push(c);
854 a_iter.next();
855 }
856
857 while let Some(&c) = b_iter.peek() {
858 if !c.is_ascii_digit() {
859 break;
860 }
861
862 b_num_str.push(c);
863 b_iter.next();
864 }
865
866 // First compare lengths (handle leading zeros)
867 match a_num_str.len().cmp(&b_num_str.len()) {
868 Ordering::Equal => {
869 // Same length, compare digit by digit
870 match a_num_str.cmp(&b_num_str) {
871 Ordering::Equal => Ordering::Equal,
872 ordering => ordering,
873 }
874 }
875
876 // Different lengths but same value means leading zeros
877 ordering => {
878 // Try parsing as numbers first
879 if let (Ok(a_val), Ok(b_val)) = (a_num_str.parse::<u128>(), b_num_str.parse::<u128>()) {
880 match a_val.cmp(&b_val) {
881 Ordering::Equal => ordering, // Same value, longer one is greater (leading zeros)
882 ord => ord,
883 }
884 } else {
885 // If parsing fails (overflow), compare as strings
886 a_num_str.cmp(&b_num_str)
887 }
888 }
889 }
890}
891
892/// Performs natural sorting comparison between two strings.
893///
894/// Natural sorting is an ordering that handles numeric sequences in a way that matches human expectations.
895/// For example, "file2" comes before "file10" (unlike standard lexicographic sorting).
896///
897/// # Characteristics
898///
899/// * Case-sensitive with lowercase priority: When comparing same letters, lowercase comes before uppercase
900/// * Numbers are compared by numeric value, not character by character
901/// * Leading zeros affect ordering when numeric values are equal
902/// * Can handle numbers larger than u128::MAX (falls back to string comparison)
903/// * When strings are equal case-insensitively, lowercase is prioritized (lowercase < uppercase)
904///
905/// # Algorithm
906///
907/// The function works by:
908/// 1. Processing strings character by character in a case-insensitive manner
909/// 2. When encountering digits, treating consecutive digits as a single number
910/// 3. Comparing numbers by their numeric value rather than lexicographically
911/// 4. For non-numeric characters, using case-insensitive comparison
912/// 5. If everything is equal case-insensitively, using case-sensitive comparison as final tie-breaker
913pub fn natural_sort(a: &str, b: &str) -> Ordering {
914 let mut a_iter = a.chars().peekable();
915 let mut b_iter = b.chars().peekable();
916
917 loop {
918 match (a_iter.peek(), b_iter.peek()) {
919 (None, None) => {
920 return b.cmp(a);
921 }
922 (None, _) => return Ordering::Less,
923 (_, None) => return Ordering::Greater,
924 (Some(&a_char), Some(&b_char)) => {
925 if a_char.is_ascii_digit() && b_char.is_ascii_digit() {
926 match compare_numeric_segments(&mut a_iter, &mut b_iter) {
927 Ordering::Equal => continue,
928 ordering => return ordering,
929 }
930 } else {
931 match a_char
932 .to_ascii_lowercase()
933 .cmp(&b_char.to_ascii_lowercase())
934 {
935 Ordering::Equal => {
936 a_iter.next();
937 b_iter.next();
938 }
939 ordering => return ordering,
940 }
941 }
942 }
943 }
944 }
945}
946
947/// Case-insensitive natural sort without applying the final lowercase/uppercase tie-breaker.
948/// This is useful when comparing individual path components where we want to keep walking
949/// deeper components before deciding on casing.
950fn natural_sort_no_tiebreak(a: &str, b: &str) -> Ordering {
951 if a.eq_ignore_ascii_case(b) {
952 Ordering::Equal
953 } else {
954 natural_sort(a, b)
955 }
956}
957
958fn stem_and_extension(filename: &str) -> (Option<&str>, Option<&str>) {
959 if filename.is_empty() {
960 return (None, None);
961 }
962
963 match filename.rsplit_once('.') {
964 // Case 1: No dot was found. The entire name is the stem.
965 None => (Some(filename), None),
966
967 // Case 2: A dot was found.
968 Some((before, after)) => {
969 // This is the crucial check for dotfiles like ".bashrc".
970 // If `before` is empty, the dot was the first character.
971 // In that case, we revert to the "whole name is the stem" logic.
972 if before.is_empty() {
973 (Some(filename), None)
974 } else {
975 // Otherwise, we have a standard stem and extension.
976 (Some(before), Some(after))
977 }
978 }
979 }
980}
981
982pub fn compare_rel_paths(
983 (path_a, a_is_file): (&RelPath, bool),
984 (path_b, b_is_file): (&RelPath, bool),
985) -> Ordering {
986 let mut components_a = path_a.components();
987 let mut components_b = path_b.components();
988 loop {
989 match (components_a.next(), components_b.next()) {
990 (Some(component_a), Some(component_b)) => {
991 let a_is_file = a_is_file && components_a.rest().is_empty();
992 let b_is_file = b_is_file && components_b.rest().is_empty();
993
994 let ordering = a_is_file.cmp(&b_is_file).then_with(|| {
995 let (a_stem, a_extension) = a_is_file
996 .then(|| stem_and_extension(component_a))
997 .unwrap_or_default();
998 let path_string_a = if a_is_file { a_stem } else { Some(component_a) };
999
1000 let (b_stem, b_extension) = b_is_file
1001 .then(|| stem_and_extension(component_b))
1002 .unwrap_or_default();
1003 let path_string_b = if b_is_file { b_stem } else { Some(component_b) };
1004
1005 let compare_components = match (path_string_a, path_string_b) {
1006 (Some(a), Some(b)) => natural_sort(&a, &b),
1007 (Some(_), None) => Ordering::Greater,
1008 (None, Some(_)) => Ordering::Less,
1009 (None, None) => Ordering::Equal,
1010 };
1011
1012 compare_components.then_with(|| {
1013 if a_is_file && b_is_file {
1014 let ext_a = a_extension.unwrap_or_default();
1015 let ext_b = b_extension.unwrap_or_default();
1016 ext_a.cmp(ext_b)
1017 } else {
1018 Ordering::Equal
1019 }
1020 })
1021 });
1022
1023 if !ordering.is_eq() {
1024 return ordering;
1025 }
1026 }
1027 (Some(_), None) => break Ordering::Greater,
1028 (None, Some(_)) => break Ordering::Less,
1029 (None, None) => break Ordering::Equal,
1030 }
1031 }
1032}
1033
1034/// Compare two relative paths with mixed files and directories using
1035/// case-insensitive natural sorting. For example, "Apple", "aardvark.txt",
1036/// and "Zebra" would be sorted as: aardvark.txt, Apple, Zebra
1037/// (case-insensitive alphabetical).
1038pub fn compare_rel_paths_mixed(
1039 (path_a, a_is_file): (&RelPath, bool),
1040 (path_b, b_is_file): (&RelPath, bool),
1041) -> Ordering {
1042 let original_paths_equal = std::ptr::eq(path_a, path_b) || path_a == path_b;
1043 let mut components_a = path_a.components();
1044 let mut components_b = path_b.components();
1045
1046 loop {
1047 match (components_a.next(), components_b.next()) {
1048 (Some(component_a), Some(component_b)) => {
1049 let a_leaf_file = a_is_file && components_a.rest().is_empty();
1050 let b_leaf_file = b_is_file && components_b.rest().is_empty();
1051
1052 let (a_stem, a_ext) = a_leaf_file
1053 .then(|| stem_and_extension(component_a))
1054 .unwrap_or_default();
1055 let (b_stem, b_ext) = b_leaf_file
1056 .then(|| stem_and_extension(component_b))
1057 .unwrap_or_default();
1058 let a_key = if a_leaf_file {
1059 a_stem
1060 } else {
1061 Some(component_a)
1062 };
1063 let b_key = if b_leaf_file {
1064 b_stem
1065 } else {
1066 Some(component_b)
1067 };
1068
1069 let ordering = match (a_key, b_key) {
1070 (Some(a), Some(b)) => natural_sort_no_tiebreak(a, b)
1071 .then_with(|| match (a_leaf_file, b_leaf_file) {
1072 (true, false) if a == b => Ordering::Greater,
1073 (false, true) if a == b => Ordering::Less,
1074 _ => Ordering::Equal,
1075 })
1076 .then_with(|| {
1077 if a_leaf_file && b_leaf_file {
1078 let a_ext_str = a_ext.unwrap_or_default().to_lowercase();
1079 let b_ext_str = b_ext.unwrap_or_default().to_lowercase();
1080 b_ext_str.cmp(&a_ext_str)
1081 } else {
1082 Ordering::Equal
1083 }
1084 }),
1085 (Some(_), None) => Ordering::Greater,
1086 (None, Some(_)) => Ordering::Less,
1087 (None, None) => Ordering::Equal,
1088 };
1089
1090 if !ordering.is_eq() {
1091 return ordering;
1092 }
1093 }
1094 (Some(_), None) => return Ordering::Greater,
1095 (None, Some(_)) => return Ordering::Less,
1096 (None, None) => {
1097 // Deterministic tie-break: use natural sort to prefer lowercase when paths
1098 // are otherwise equal but still differ in casing.
1099 if !original_paths_equal {
1100 return natural_sort(path_a.as_unix_str(), path_b.as_unix_str());
1101 }
1102 return Ordering::Equal;
1103 }
1104 }
1105 }
1106}
1107
1108/// Compare two relative paths with files before directories using
1109/// case-insensitive natural sorting. At each directory level, all files
1110/// are sorted before all directories, with case-insensitive alphabetical
1111/// ordering within each group.
1112pub fn compare_rel_paths_files_first(
1113 (path_a, a_is_file): (&RelPath, bool),
1114 (path_b, b_is_file): (&RelPath, bool),
1115) -> Ordering {
1116 let original_paths_equal = std::ptr::eq(path_a, path_b) || path_a == path_b;
1117 let mut components_a = path_a.components();
1118 let mut components_b = path_b.components();
1119
1120 loop {
1121 match (components_a.next(), components_b.next()) {
1122 (Some(component_a), Some(component_b)) => {
1123 let a_leaf_file = a_is_file && components_a.rest().is_empty();
1124 let b_leaf_file = b_is_file && components_b.rest().is_empty();
1125
1126 let (a_stem, a_ext) = a_leaf_file
1127 .then(|| stem_and_extension(component_a))
1128 .unwrap_or_default();
1129 let (b_stem, b_ext) = b_leaf_file
1130 .then(|| stem_and_extension(component_b))
1131 .unwrap_or_default();
1132 let a_key = if a_leaf_file {
1133 a_stem
1134 } else {
1135 Some(component_a)
1136 };
1137 let b_key = if b_leaf_file {
1138 b_stem
1139 } else {
1140 Some(component_b)
1141 };
1142
1143 let ordering = match (a_key, b_key) {
1144 (Some(a), Some(b)) => {
1145 if a_leaf_file && !b_leaf_file {
1146 Ordering::Less
1147 } else if !a_leaf_file && b_leaf_file {
1148 Ordering::Greater
1149 } else {
1150 natural_sort_no_tiebreak(a, b).then_with(|| {
1151 if a_leaf_file && b_leaf_file {
1152 let a_ext_str = a_ext.unwrap_or_default().to_lowercase();
1153 let b_ext_str = b_ext.unwrap_or_default().to_lowercase();
1154 a_ext_str.cmp(&b_ext_str)
1155 } else {
1156 Ordering::Equal
1157 }
1158 })
1159 }
1160 }
1161 (Some(_), None) => Ordering::Greater,
1162 (None, Some(_)) => Ordering::Less,
1163 (None, None) => Ordering::Equal,
1164 };
1165
1166 if !ordering.is_eq() {
1167 return ordering;
1168 }
1169 }
1170 (Some(_), None) => return Ordering::Greater,
1171 (None, Some(_)) => return Ordering::Less,
1172 (None, None) => {
1173 // Deterministic tie-break: use natural sort to prefer lowercase when paths
1174 // are otherwise equal but still differ in casing.
1175 if !original_paths_equal {
1176 return natural_sort(path_a.as_unix_str(), path_b.as_unix_str());
1177 }
1178 return Ordering::Equal;
1179 }
1180 }
1181 }
1182}
1183
1184pub fn compare_paths(
1185 (path_a, a_is_file): (&Path, bool),
1186 (path_b, b_is_file): (&Path, bool),
1187) -> Ordering {
1188 let mut components_a = path_a.components().peekable();
1189 let mut components_b = path_b.components().peekable();
1190
1191 loop {
1192 match (components_a.next(), components_b.next()) {
1193 (Some(component_a), Some(component_b)) => {
1194 let a_is_file = components_a.peek().is_none() && a_is_file;
1195 let b_is_file = components_b.peek().is_none() && b_is_file;
1196
1197 let ordering = a_is_file.cmp(&b_is_file).then_with(|| {
1198 let path_a = Path::new(component_a.as_os_str());
1199 let path_string_a = if a_is_file {
1200 path_a.file_stem()
1201 } else {
1202 path_a.file_name()
1203 }
1204 .map(|s| s.to_string_lossy());
1205
1206 let path_b = Path::new(component_b.as_os_str());
1207 let path_string_b = if b_is_file {
1208 path_b.file_stem()
1209 } else {
1210 path_b.file_name()
1211 }
1212 .map(|s| s.to_string_lossy());
1213
1214 let compare_components = match (path_string_a, path_string_b) {
1215 (Some(a), Some(b)) => natural_sort(&a, &b),
1216 (Some(_), None) => Ordering::Greater,
1217 (None, Some(_)) => Ordering::Less,
1218 (None, None) => Ordering::Equal,
1219 };
1220
1221 compare_components.then_with(|| {
1222 if a_is_file && b_is_file {
1223 let ext_a = path_a.extension().unwrap_or_default();
1224 let ext_b = path_b.extension().unwrap_or_default();
1225 ext_a.cmp(ext_b)
1226 } else {
1227 Ordering::Equal
1228 }
1229 })
1230 });
1231
1232 if !ordering.is_eq() {
1233 return ordering;
1234 }
1235 }
1236 (Some(_), None) => break Ordering::Greater,
1237 (None, Some(_)) => break Ordering::Less,
1238 (None, None) => break Ordering::Equal,
1239 }
1240 }
1241}
1242
1243#[derive(Debug, Clone, PartialEq, Eq)]
1244pub struct WslPath {
1245 pub distro: String,
1246
1247 // the reason this is an OsString and not any of the path types is that it needs to
1248 // represent a unix path (with '/' separators) on windows. `from_path` does this by
1249 // manually constructing it from the path components of a given windows path.
1250 pub path: std::ffi::OsString,
1251}
1252
1253impl WslPath {
1254 pub fn from_path<P: AsRef<Path>>(path: P) -> Option<WslPath> {
1255 if cfg!(not(target_os = "windows")) {
1256 return None;
1257 }
1258 use std::{
1259 ffi::OsString,
1260 path::{Component, Prefix},
1261 };
1262
1263 let mut components = path.as_ref().components();
1264 let Some(Component::Prefix(prefix)) = components.next() else {
1265 return None;
1266 };
1267 let (server, distro) = match prefix.kind() {
1268 Prefix::UNC(server, distro) => (server, distro),
1269 Prefix::VerbatimUNC(server, distro) => (server, distro),
1270 _ => return None,
1271 };
1272 let Some(Component::RootDir) = components.next() else {
1273 return None;
1274 };
1275
1276 let server_str = server.to_string_lossy();
1277 if server_str == "wsl.localhost" || server_str == "wsl$" {
1278 let mut result = OsString::from("");
1279 for c in components {
1280 use Component::*;
1281 match c {
1282 Prefix(p) => unreachable!("got {p:?}, but already stripped prefix"),
1283 RootDir => unreachable!("got root dir, but already stripped root"),
1284 CurDir => continue,
1285 ParentDir => result.push("/.."),
1286 Normal(s) => {
1287 result.push("/");
1288 result.push(s);
1289 }
1290 }
1291 }
1292 if result.is_empty() {
1293 result.push("/");
1294 }
1295 Some(WslPath {
1296 distro: distro.to_string_lossy().to_string(),
1297 path: result,
1298 })
1299 } else {
1300 None
1301 }
1302 }
1303}
1304
1305#[cfg(test)]
1306mod tests {
1307 use super::*;
1308 use util_macros::perf;
1309
1310 #[perf]
1311 fn compare_paths_with_dots() {
1312 let mut paths = vec![
1313 (Path::new("test_dirs"), false),
1314 (Path::new("test_dirs/1.46"), false),
1315 (Path::new("test_dirs/1.46/bar_1"), true),
1316 (Path::new("test_dirs/1.46/bar_2"), true),
1317 (Path::new("test_dirs/1.45"), false),
1318 (Path::new("test_dirs/1.45/foo_2"), true),
1319 (Path::new("test_dirs/1.45/foo_1"), true),
1320 ];
1321 paths.sort_by(|&a, &b| compare_paths(a, b));
1322 assert_eq!(
1323 paths,
1324 vec![
1325 (Path::new("test_dirs"), false),
1326 (Path::new("test_dirs/1.45"), false),
1327 (Path::new("test_dirs/1.45/foo_1"), true),
1328 (Path::new("test_dirs/1.45/foo_2"), true),
1329 (Path::new("test_dirs/1.46"), false),
1330 (Path::new("test_dirs/1.46/bar_1"), true),
1331 (Path::new("test_dirs/1.46/bar_2"), true),
1332 ]
1333 );
1334 let mut paths = vec![
1335 (Path::new("root1/one.txt"), true),
1336 (Path::new("root1/one.two.txt"), true),
1337 ];
1338 paths.sort_by(|&a, &b| compare_paths(a, b));
1339 assert_eq!(
1340 paths,
1341 vec![
1342 (Path::new("root1/one.txt"), true),
1343 (Path::new("root1/one.two.txt"), true),
1344 ]
1345 );
1346 }
1347
1348 #[perf]
1349 fn compare_paths_with_same_name_different_extensions() {
1350 let mut paths = vec![
1351 (Path::new("test_dirs/file.rs"), true),
1352 (Path::new("test_dirs/file.txt"), true),
1353 (Path::new("test_dirs/file.md"), true),
1354 (Path::new("test_dirs/file"), true),
1355 (Path::new("test_dirs/file.a"), true),
1356 ];
1357 paths.sort_by(|&a, &b| compare_paths(a, b));
1358 assert_eq!(
1359 paths,
1360 vec![
1361 (Path::new("test_dirs/file"), true),
1362 (Path::new("test_dirs/file.a"), true),
1363 (Path::new("test_dirs/file.md"), true),
1364 (Path::new("test_dirs/file.rs"), true),
1365 (Path::new("test_dirs/file.txt"), true),
1366 ]
1367 );
1368 }
1369
1370 #[perf]
1371 fn compare_paths_case_semi_sensitive() {
1372 let mut paths = vec![
1373 (Path::new("test_DIRS"), false),
1374 (Path::new("test_DIRS/foo_1"), true),
1375 (Path::new("test_DIRS/foo_2"), true),
1376 (Path::new("test_DIRS/bar"), true),
1377 (Path::new("test_DIRS/BAR"), true),
1378 (Path::new("test_dirs"), false),
1379 (Path::new("test_dirs/foo_1"), true),
1380 (Path::new("test_dirs/foo_2"), true),
1381 (Path::new("test_dirs/bar"), true),
1382 (Path::new("test_dirs/BAR"), true),
1383 ];
1384 paths.sort_by(|&a, &b| compare_paths(a, b));
1385 assert_eq!(
1386 paths,
1387 vec![
1388 (Path::new("test_dirs"), false),
1389 (Path::new("test_dirs/bar"), true),
1390 (Path::new("test_dirs/BAR"), true),
1391 (Path::new("test_dirs/foo_1"), true),
1392 (Path::new("test_dirs/foo_2"), true),
1393 (Path::new("test_DIRS"), false),
1394 (Path::new("test_DIRS/bar"), true),
1395 (Path::new("test_DIRS/BAR"), true),
1396 (Path::new("test_DIRS/foo_1"), true),
1397 (Path::new("test_DIRS/foo_2"), true),
1398 ]
1399 );
1400 }
1401
1402 #[perf]
1403 fn compare_paths_mixed_case_numeric_ordering() {
1404 let mut entries = [
1405 (Path::new(".config"), false),
1406 (Path::new("Dir1"), false),
1407 (Path::new("dir01"), false),
1408 (Path::new("dir2"), false),
1409 (Path::new("Dir02"), false),
1410 (Path::new("dir10"), false),
1411 (Path::new("Dir10"), false),
1412 ];
1413
1414 entries.sort_by(|&a, &b| compare_paths(a, b));
1415
1416 let ordered: Vec<&str> = entries
1417 .iter()
1418 .map(|(path, _)| path.to_str().unwrap())
1419 .collect();
1420
1421 assert_eq!(
1422 ordered,
1423 vec![
1424 ".config", "Dir1", "dir01", "dir2", "Dir02", "dir10", "Dir10"
1425 ]
1426 );
1427 }
1428
1429 #[perf]
1430 fn compare_rel_paths_mixed_case_insensitive() {
1431 // Test that mixed mode is case-insensitive
1432 let mut paths = vec![
1433 (RelPath::unix("zebra.txt").unwrap(), true),
1434 (RelPath::unix("Apple").unwrap(), false),
1435 (RelPath::unix("banana.rs").unwrap(), true),
1436 (RelPath::unix("Carrot").unwrap(), false),
1437 (RelPath::unix("aardvark.txt").unwrap(), true),
1438 ];
1439 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1440 // Case-insensitive: aardvark < Apple < banana < Carrot < zebra
1441 assert_eq!(
1442 paths,
1443 vec![
1444 (RelPath::unix("aardvark.txt").unwrap(), true),
1445 (RelPath::unix("Apple").unwrap(), false),
1446 (RelPath::unix("banana.rs").unwrap(), true),
1447 (RelPath::unix("Carrot").unwrap(), false),
1448 (RelPath::unix("zebra.txt").unwrap(), true),
1449 ]
1450 );
1451 }
1452
1453 #[perf]
1454 fn compare_rel_paths_files_first_basic() {
1455 // Test that files come before directories
1456 let mut paths = vec![
1457 (RelPath::unix("zebra.txt").unwrap(), true),
1458 (RelPath::unix("Apple").unwrap(), false),
1459 (RelPath::unix("banana.rs").unwrap(), true),
1460 (RelPath::unix("Carrot").unwrap(), false),
1461 (RelPath::unix("aardvark.txt").unwrap(), true),
1462 ];
1463 paths.sort_by(|&a, &b| compare_rel_paths_files_first(a, b));
1464 // Files first (case-insensitive), then directories (case-insensitive)
1465 assert_eq!(
1466 paths,
1467 vec![
1468 (RelPath::unix("aardvark.txt").unwrap(), true),
1469 (RelPath::unix("banana.rs").unwrap(), true),
1470 (RelPath::unix("zebra.txt").unwrap(), true),
1471 (RelPath::unix("Apple").unwrap(), false),
1472 (RelPath::unix("Carrot").unwrap(), false),
1473 ]
1474 );
1475 }
1476
1477 #[perf]
1478 fn compare_rel_paths_files_first_case_insensitive() {
1479 // Test case-insensitive sorting within files and directories
1480 let mut paths = vec![
1481 (RelPath::unix("Zebra.txt").unwrap(), true),
1482 (RelPath::unix("apple").unwrap(), false),
1483 (RelPath::unix("Banana.rs").unwrap(), true),
1484 (RelPath::unix("carrot").unwrap(), false),
1485 (RelPath::unix("Aardvark.txt").unwrap(), true),
1486 ];
1487 paths.sort_by(|&a, &b| compare_rel_paths_files_first(a, b));
1488 assert_eq!(
1489 paths,
1490 vec![
1491 (RelPath::unix("Aardvark.txt").unwrap(), true),
1492 (RelPath::unix("Banana.rs").unwrap(), true),
1493 (RelPath::unix("Zebra.txt").unwrap(), true),
1494 (RelPath::unix("apple").unwrap(), false),
1495 (RelPath::unix("carrot").unwrap(), false),
1496 ]
1497 );
1498 }
1499
1500 #[perf]
1501 fn compare_rel_paths_files_first_numeric() {
1502 // Test natural number sorting with files first
1503 let mut paths = vec![
1504 (RelPath::unix("file10.txt").unwrap(), true),
1505 (RelPath::unix("dir2").unwrap(), false),
1506 (RelPath::unix("file2.txt").unwrap(), true),
1507 (RelPath::unix("dir10").unwrap(), false),
1508 (RelPath::unix("file1.txt").unwrap(), true),
1509 ];
1510 paths.sort_by(|&a, &b| compare_rel_paths_files_first(a, b));
1511 assert_eq!(
1512 paths,
1513 vec![
1514 (RelPath::unix("file1.txt").unwrap(), true),
1515 (RelPath::unix("file2.txt").unwrap(), true),
1516 (RelPath::unix("file10.txt").unwrap(), true),
1517 (RelPath::unix("dir2").unwrap(), false),
1518 (RelPath::unix("dir10").unwrap(), false),
1519 ]
1520 );
1521 }
1522
1523 #[perf]
1524 fn compare_rel_paths_mixed_case() {
1525 // Test case-insensitive sorting with varied capitalization
1526 let mut paths = vec![
1527 (RelPath::unix("README.md").unwrap(), true),
1528 (RelPath::unix("readme.txt").unwrap(), true),
1529 (RelPath::unix("ReadMe.rs").unwrap(), true),
1530 ];
1531 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1532 // All "readme" variants should group together, sorted by extension
1533 assert_eq!(
1534 paths,
1535 vec![
1536 (RelPath::unix("readme.txt").unwrap(), true),
1537 (RelPath::unix("ReadMe.rs").unwrap(), true),
1538 (RelPath::unix("README.md").unwrap(), true),
1539 ]
1540 );
1541 }
1542
1543 #[perf]
1544 fn compare_rel_paths_mixed_files_and_dirs() {
1545 // Verify directories and files are still mixed
1546 let mut paths = vec![
1547 (RelPath::unix("file2.txt").unwrap(), true),
1548 (RelPath::unix("Dir1").unwrap(), false),
1549 (RelPath::unix("file1.txt").unwrap(), true),
1550 (RelPath::unix("dir2").unwrap(), false),
1551 ];
1552 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1553 // Case-insensitive: dir1, dir2, file1, file2 (all mixed)
1554 assert_eq!(
1555 paths,
1556 vec![
1557 (RelPath::unix("Dir1").unwrap(), false),
1558 (RelPath::unix("dir2").unwrap(), false),
1559 (RelPath::unix("file1.txt").unwrap(), true),
1560 (RelPath::unix("file2.txt").unwrap(), true),
1561 ]
1562 );
1563 }
1564
1565 #[perf]
1566 fn compare_rel_paths_mixed_with_nested_paths() {
1567 // Test that nested paths still work correctly
1568 let mut paths = vec![
1569 (RelPath::unix("src/main.rs").unwrap(), true),
1570 (RelPath::unix("Cargo.toml").unwrap(), true),
1571 (RelPath::unix("src").unwrap(), false),
1572 (RelPath::unix("target").unwrap(), false),
1573 ];
1574 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1575 assert_eq!(
1576 paths,
1577 vec![
1578 (RelPath::unix("Cargo.toml").unwrap(), true),
1579 (RelPath::unix("src").unwrap(), false),
1580 (RelPath::unix("src/main.rs").unwrap(), true),
1581 (RelPath::unix("target").unwrap(), false),
1582 ]
1583 );
1584 }
1585
1586 #[perf]
1587 fn compare_rel_paths_files_first_with_nested() {
1588 // Files come before directories, even with nested paths
1589 let mut paths = vec![
1590 (RelPath::unix("src/lib.rs").unwrap(), true),
1591 (RelPath::unix("README.md").unwrap(), true),
1592 (RelPath::unix("src").unwrap(), false),
1593 (RelPath::unix("tests").unwrap(), false),
1594 ];
1595 paths.sort_by(|&a, &b| compare_rel_paths_files_first(a, b));
1596 assert_eq!(
1597 paths,
1598 vec![
1599 (RelPath::unix("README.md").unwrap(), true),
1600 (RelPath::unix("src").unwrap(), false),
1601 (RelPath::unix("src/lib.rs").unwrap(), true),
1602 (RelPath::unix("tests").unwrap(), false),
1603 ]
1604 );
1605 }
1606
1607 #[perf]
1608 fn compare_rel_paths_mixed_dotfiles() {
1609 // Test that dotfiles are handled correctly in mixed mode
1610 let mut paths = vec![
1611 (RelPath::unix(".gitignore").unwrap(), true),
1612 (RelPath::unix("README.md").unwrap(), true),
1613 (RelPath::unix(".github").unwrap(), false),
1614 (RelPath::unix("src").unwrap(), false),
1615 ];
1616 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1617 assert_eq!(
1618 paths,
1619 vec![
1620 (RelPath::unix(".github").unwrap(), false),
1621 (RelPath::unix(".gitignore").unwrap(), true),
1622 (RelPath::unix("README.md").unwrap(), true),
1623 (RelPath::unix("src").unwrap(), false),
1624 ]
1625 );
1626 }
1627
1628 #[perf]
1629 fn compare_rel_paths_files_first_dotfiles() {
1630 // Test that dotfiles come first when they're files
1631 let mut paths = vec![
1632 (RelPath::unix(".gitignore").unwrap(), true),
1633 (RelPath::unix("README.md").unwrap(), true),
1634 (RelPath::unix(".github").unwrap(), false),
1635 (RelPath::unix("src").unwrap(), false),
1636 ];
1637 paths.sort_by(|&a, &b| compare_rel_paths_files_first(a, b));
1638 assert_eq!(
1639 paths,
1640 vec![
1641 (RelPath::unix(".gitignore").unwrap(), true),
1642 (RelPath::unix("README.md").unwrap(), true),
1643 (RelPath::unix(".github").unwrap(), false),
1644 (RelPath::unix("src").unwrap(), false),
1645 ]
1646 );
1647 }
1648
1649 #[perf]
1650 fn compare_rel_paths_mixed_same_stem_different_extension() {
1651 // Files with same stem but different extensions should sort by extension
1652 let mut paths = vec![
1653 (RelPath::unix("file.rs").unwrap(), true),
1654 (RelPath::unix("file.md").unwrap(), true),
1655 (RelPath::unix("file.txt").unwrap(), true),
1656 ];
1657 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1658 assert_eq!(
1659 paths,
1660 vec![
1661 (RelPath::unix("file.txt").unwrap(), true),
1662 (RelPath::unix("file.rs").unwrap(), true),
1663 (RelPath::unix("file.md").unwrap(), true),
1664 ]
1665 );
1666 }
1667
1668 #[perf]
1669 fn compare_rel_paths_files_first_same_stem() {
1670 // Same stem files should still sort by extension with files_first
1671 let mut paths = vec![
1672 (RelPath::unix("main.rs").unwrap(), true),
1673 (RelPath::unix("main.c").unwrap(), true),
1674 (RelPath::unix("main").unwrap(), false),
1675 ];
1676 paths.sort_by(|&a, &b| compare_rel_paths_files_first(a, b));
1677 assert_eq!(
1678 paths,
1679 vec![
1680 (RelPath::unix("main.c").unwrap(), true),
1681 (RelPath::unix("main.rs").unwrap(), true),
1682 (RelPath::unix("main").unwrap(), false),
1683 ]
1684 );
1685 }
1686
1687 #[perf]
1688 fn compare_rel_paths_mixed_deep_nesting() {
1689 // Test sorting with deeply nested paths
1690 let mut paths = vec![
1691 (RelPath::unix("a/b/c.txt").unwrap(), true),
1692 (RelPath::unix("A/B.txt").unwrap(), true),
1693 (RelPath::unix("a.txt").unwrap(), true),
1694 (RelPath::unix("A.txt").unwrap(), true),
1695 ];
1696 paths.sort_by(|&a, &b| compare_rel_paths_mixed(a, b));
1697 assert_eq!(
1698 paths,
1699 vec![
1700 (RelPath::unix("A/B.txt").unwrap(), true),
1701 (RelPath::unix("a/b/c.txt").unwrap(), true),
1702 (RelPath::unix("a.txt").unwrap(), true),
1703 (RelPath::unix("A.txt").unwrap(), true),
1704 ]
1705 );
1706 }
1707
1708 #[perf]
1709 fn path_with_position_parse_posix_path() {
1710 // Test POSIX filename edge cases
1711 // Read more at https://en.wikipedia.org/wiki/Filename
1712 assert_eq!(
1713 PathWithPosition::parse_str("test_file"),
1714 PathWithPosition {
1715 path: PathBuf::from("test_file"),
1716 row: None,
1717 column: None
1718 }
1719 );
1720
1721 assert_eq!(
1722 PathWithPosition::parse_str("a:bc:.zip:1"),
1723 PathWithPosition {
1724 path: PathBuf::from("a:bc:.zip"),
1725 row: Some(1),
1726 column: None
1727 }
1728 );
1729
1730 assert_eq!(
1731 PathWithPosition::parse_str("one.second.zip:1"),
1732 PathWithPosition {
1733 path: PathBuf::from("one.second.zip"),
1734 row: Some(1),
1735 column: None
1736 }
1737 );
1738
1739 // Trim off trailing `:`s for otherwise valid input.
1740 assert_eq!(
1741 PathWithPosition::parse_str("test_file:10:1:"),
1742 PathWithPosition {
1743 path: PathBuf::from("test_file"),
1744 row: Some(10),
1745 column: Some(1)
1746 }
1747 );
1748
1749 assert_eq!(
1750 PathWithPosition::parse_str("test_file.rs:"),
1751 PathWithPosition {
1752 path: PathBuf::from("test_file.rs"),
1753 row: None,
1754 column: None
1755 }
1756 );
1757
1758 assert_eq!(
1759 PathWithPosition::parse_str("test_file.rs:1:"),
1760 PathWithPosition {
1761 path: PathBuf::from("test_file.rs"),
1762 row: Some(1),
1763 column: None
1764 }
1765 );
1766
1767 assert_eq!(
1768 PathWithPosition::parse_str("ab\ncd"),
1769 PathWithPosition {
1770 path: PathBuf::from("ab\ncd"),
1771 row: None,
1772 column: None
1773 }
1774 );
1775
1776 assert_eq!(
1777 PathWithPosition::parse_str("👋\nab"),
1778 PathWithPosition {
1779 path: PathBuf::from("👋\nab"),
1780 row: None,
1781 column: None
1782 }
1783 );
1784
1785 assert_eq!(
1786 PathWithPosition::parse_str("Types.hs:(617,9)-(670,28):"),
1787 PathWithPosition {
1788 path: PathBuf::from("Types.hs"),
1789 row: Some(617),
1790 column: Some(9),
1791 }
1792 );
1793 }
1794
1795 #[perf]
1796 #[cfg(not(target_os = "windows"))]
1797 fn path_with_position_parse_posix_path_with_suffix() {
1798 assert_eq!(
1799 PathWithPosition::parse_str("foo/bar:34:in"),
1800 PathWithPosition {
1801 path: PathBuf::from("foo/bar"),
1802 row: Some(34),
1803 column: None,
1804 }
1805 );
1806 assert_eq!(
1807 PathWithPosition::parse_str("foo/bar.rs:1902:::15:"),
1808 PathWithPosition {
1809 path: PathBuf::from("foo/bar.rs:1902"),
1810 row: Some(15),
1811 column: None
1812 }
1813 );
1814
1815 assert_eq!(
1816 PathWithPosition::parse_str("app-editors:zed-0.143.6:20240710-201212.log:34:"),
1817 PathWithPosition {
1818 path: PathBuf::from("app-editors:zed-0.143.6:20240710-201212.log"),
1819 row: Some(34),
1820 column: None,
1821 }
1822 );
1823
1824 assert_eq!(
1825 PathWithPosition::parse_str("crates/file_finder/src/file_finder.rs:1902:13:"),
1826 PathWithPosition {
1827 path: PathBuf::from("crates/file_finder/src/file_finder.rs"),
1828 row: Some(1902),
1829 column: Some(13),
1830 }
1831 );
1832
1833 assert_eq!(
1834 PathWithPosition::parse_str("crate/utils/src/test:today.log:34"),
1835 PathWithPosition {
1836 path: PathBuf::from("crate/utils/src/test:today.log"),
1837 row: Some(34),
1838 column: None,
1839 }
1840 );
1841 assert_eq!(
1842 PathWithPosition::parse_str("/testing/out/src/file_finder.odin(7:15)"),
1843 PathWithPosition {
1844 path: PathBuf::from("/testing/out/src/file_finder.odin"),
1845 row: Some(7),
1846 column: Some(15),
1847 }
1848 );
1849 }
1850
1851 #[perf]
1852 #[cfg(target_os = "windows")]
1853 fn path_with_position_parse_windows_path() {
1854 assert_eq!(
1855 PathWithPosition::parse_str("crates\\utils\\paths.rs"),
1856 PathWithPosition {
1857 path: PathBuf::from("crates\\utils\\paths.rs"),
1858 row: None,
1859 column: None
1860 }
1861 );
1862
1863 assert_eq!(
1864 PathWithPosition::parse_str("C:\\Users\\someone\\test_file.rs"),
1865 PathWithPosition {
1866 path: PathBuf::from("C:\\Users\\someone\\test_file.rs"),
1867 row: None,
1868 column: None
1869 }
1870 );
1871 }
1872
1873 #[perf]
1874 #[cfg(target_os = "windows")]
1875 fn path_with_position_parse_windows_path_with_suffix() {
1876 assert_eq!(
1877 PathWithPosition::parse_str("crates\\utils\\paths.rs:101"),
1878 PathWithPosition {
1879 path: PathBuf::from("crates\\utils\\paths.rs"),
1880 row: Some(101),
1881 column: None
1882 }
1883 );
1884
1885 assert_eq!(
1886 PathWithPosition::parse_str("\\\\?\\C:\\Users\\someone\\test_file.rs:1:20"),
1887 PathWithPosition {
1888 path: PathBuf::from("\\\\?\\C:\\Users\\someone\\test_file.rs"),
1889 row: Some(1),
1890 column: Some(20)
1891 }
1892 );
1893
1894 assert_eq!(
1895 PathWithPosition::parse_str("C:\\Users\\someone\\test_file.rs(1902,13)"),
1896 PathWithPosition {
1897 path: PathBuf::from("C:\\Users\\someone\\test_file.rs"),
1898 row: Some(1902),
1899 column: Some(13)
1900 }
1901 );
1902
1903 // Trim off trailing `:`s for otherwise valid input.
1904 assert_eq!(
1905 PathWithPosition::parse_str("\\\\?\\C:\\Users\\someone\\test_file.rs:1902:13:"),
1906 PathWithPosition {
1907 path: PathBuf::from("\\\\?\\C:\\Users\\someone\\test_file.rs"),
1908 row: Some(1902),
1909 column: Some(13)
1910 }
1911 );
1912
1913 assert_eq!(
1914 PathWithPosition::parse_str("\\\\?\\C:\\Users\\someone\\test_file.rs:1902:13:15:"),
1915 PathWithPosition {
1916 path: PathBuf::from("\\\\?\\C:\\Users\\someone\\test_file.rs:1902"),
1917 row: Some(13),
1918 column: Some(15)
1919 }
1920 );
1921
1922 assert_eq!(
1923 PathWithPosition::parse_str("\\\\?\\C:\\Users\\someone\\test_file.rs:1902:::15:"),
1924 PathWithPosition {
1925 path: PathBuf::from("\\\\?\\C:\\Users\\someone\\test_file.rs:1902"),
1926 row: Some(15),
1927 column: None
1928 }
1929 );
1930
1931 assert_eq!(
1932 PathWithPosition::parse_str("\\\\?\\C:\\Users\\someone\\test_file.rs(1902,13):"),
1933 PathWithPosition {
1934 path: PathBuf::from("\\\\?\\C:\\Users\\someone\\test_file.rs"),
1935 row: Some(1902),
1936 column: Some(13),
1937 }
1938 );
1939
1940 assert_eq!(
1941 PathWithPosition::parse_str("\\\\?\\C:\\Users\\someone\\test_file.rs(1902):"),
1942 PathWithPosition {
1943 path: PathBuf::from("\\\\?\\C:\\Users\\someone\\test_file.rs"),
1944 row: Some(1902),
1945 column: None,
1946 }
1947 );
1948
1949 assert_eq!(
1950 PathWithPosition::parse_str("C:\\Users\\someone\\test_file.rs:1902:13:"),
1951 PathWithPosition {
1952 path: PathBuf::from("C:\\Users\\someone\\test_file.rs"),
1953 row: Some(1902),
1954 column: Some(13),
1955 }
1956 );
1957
1958 assert_eq!(
1959 PathWithPosition::parse_str("C:\\Users\\someone\\test_file.rs(1902,13):"),
1960 PathWithPosition {
1961 path: PathBuf::from("C:\\Users\\someone\\test_file.rs"),
1962 row: Some(1902),
1963 column: Some(13),
1964 }
1965 );
1966
1967 assert_eq!(
1968 PathWithPosition::parse_str("C:\\Users\\someone\\test_file.rs(1902):"),
1969 PathWithPosition {
1970 path: PathBuf::from("C:\\Users\\someone\\test_file.rs"),
1971 row: Some(1902),
1972 column: None,
1973 }
1974 );
1975
1976 assert_eq!(
1977 PathWithPosition::parse_str("crates/utils/paths.rs:101"),
1978 PathWithPosition {
1979 path: PathBuf::from("crates\\utils\\paths.rs"),
1980 row: Some(101),
1981 column: None,
1982 }
1983 );
1984 }
1985
1986 #[perf]
1987 fn test_path_compact() {
1988 let path: PathBuf = [
1989 home_dir().to_string_lossy().into_owned(),
1990 "some_file.txt".to_string(),
1991 ]
1992 .iter()
1993 .collect();
1994 if cfg!(any(target_os = "linux", target_os = "freebsd")) || cfg!(target_os = "macos") {
1995 assert_eq!(path.compact().to_str(), Some("~/some_file.txt"));
1996 } else {
1997 assert_eq!(path.compact().to_str(), path.to_str());
1998 }
1999 }
2000
2001 #[perf]
2002 fn test_extension_or_hidden_file_name() {
2003 // No dots in name
2004 let path = Path::new("/a/b/c/file_name.rs");
2005 assert_eq!(path.extension_or_hidden_file_name(), Some("rs"));
2006
2007 // Single dot in name
2008 let path = Path::new("/a/b/c/file.name.rs");
2009 assert_eq!(path.extension_or_hidden_file_name(), Some("rs"));
2010
2011 // Multiple dots in name
2012 let path = Path::new("/a/b/c/long.file.name.rs");
2013 assert_eq!(path.extension_or_hidden_file_name(), Some("rs"));
2014
2015 // Hidden file, no extension
2016 let path = Path::new("/a/b/c/.gitignore");
2017 assert_eq!(path.extension_or_hidden_file_name(), Some("gitignore"));
2018
2019 // Hidden file, with extension
2020 let path = Path::new("/a/b/c/.eslintrc.js");
2021 assert_eq!(path.extension_or_hidden_file_name(), Some("eslintrc.js"));
2022 }
2023
2024 #[perf]
2025 fn edge_of_glob() {
2026 let path = Path::new("/work/node_modules");
2027 let path_matcher =
2028 PathMatcher::new(&["**/node_modules/**".to_owned()], PathStyle::Posix).unwrap();
2029 assert!(
2030 path_matcher.is_match(path),
2031 "Path matcher should match {path:?}"
2032 );
2033 }
2034
2035 #[perf]
2036 fn file_in_dirs() {
2037 let path = Path::new("/work/.env");
2038 let path_matcher = PathMatcher::new(&["**/.env".to_owned()], PathStyle::Posix).unwrap();
2039 assert!(
2040 path_matcher.is_match(path),
2041 "Path matcher should match {path:?}"
2042 );
2043 let path = Path::new("/work/package.json");
2044 assert!(
2045 !path_matcher.is_match(path),
2046 "Path matcher should not match {path:?}"
2047 );
2048 }
2049
2050 #[perf]
2051 fn project_search() {
2052 let path = Path::new("/Users/someonetoignore/work/zed/zed.dev/node_modules");
2053 let path_matcher =
2054 PathMatcher::new(&["**/node_modules/**".to_owned()], PathStyle::Posix).unwrap();
2055 assert!(
2056 path_matcher.is_match(path),
2057 "Path matcher should match {path:?}"
2058 );
2059 }
2060
2061 #[perf]
2062 #[cfg(target_os = "windows")]
2063 fn test_sanitized_path() {
2064 let path = Path::new("C:\\Users\\someone\\test_file.rs");
2065 let sanitized_path = SanitizedPath::new(path);
2066 assert_eq!(
2067 sanitized_path.to_string(),
2068 "C:\\Users\\someone\\test_file.rs"
2069 );
2070
2071 let path = Path::new("\\\\?\\C:\\Users\\someone\\test_file.rs");
2072 let sanitized_path = SanitizedPath::new(path);
2073 assert_eq!(
2074 sanitized_path.to_string(),
2075 "C:\\Users\\someone\\test_file.rs"
2076 );
2077 }
2078
2079 #[perf]
2080 fn test_compare_numeric_segments() {
2081 // Helper function to create peekable iterators and test
2082 fn compare(a: &str, b: &str) -> Ordering {
2083 let mut a_iter = a.chars().peekable();
2084 let mut b_iter = b.chars().peekable();
2085
2086 let result = compare_numeric_segments(&mut a_iter, &mut b_iter);
2087
2088 // Verify iterators advanced correctly
2089 assert!(
2090 !a_iter.next().is_some_and(|c| c.is_ascii_digit()),
2091 "Iterator a should have consumed all digits"
2092 );
2093 assert!(
2094 !b_iter.next().is_some_and(|c| c.is_ascii_digit()),
2095 "Iterator b should have consumed all digits"
2096 );
2097
2098 result
2099 }
2100
2101 // Basic numeric comparisons
2102 assert_eq!(compare("0", "0"), Ordering::Equal);
2103 assert_eq!(compare("1", "2"), Ordering::Less);
2104 assert_eq!(compare("9", "10"), Ordering::Less);
2105 assert_eq!(compare("10", "9"), Ordering::Greater);
2106 assert_eq!(compare("99", "100"), Ordering::Less);
2107
2108 // Leading zeros
2109 assert_eq!(compare("0", "00"), Ordering::Less);
2110 assert_eq!(compare("00", "0"), Ordering::Greater);
2111 assert_eq!(compare("01", "1"), Ordering::Greater);
2112 assert_eq!(compare("001", "1"), Ordering::Greater);
2113 assert_eq!(compare("001", "01"), Ordering::Greater);
2114
2115 // Same value different representation
2116 assert_eq!(compare("000100", "100"), Ordering::Greater);
2117 assert_eq!(compare("100", "0100"), Ordering::Less);
2118 assert_eq!(compare("0100", "00100"), Ordering::Less);
2119
2120 // Large numbers
2121 assert_eq!(compare("9999999999", "10000000000"), Ordering::Less);
2122 assert_eq!(
2123 compare(
2124 "340282366920938463463374607431768211455", // u128::MAX
2125 "340282366920938463463374607431768211456"
2126 ),
2127 Ordering::Less
2128 );
2129 assert_eq!(
2130 compare(
2131 "340282366920938463463374607431768211456", // > u128::MAX
2132 "340282366920938463463374607431768211455"
2133 ),
2134 Ordering::Greater
2135 );
2136
2137 // Iterator advancement verification
2138 let mut a_iter = "123abc".chars().peekable();
2139 let mut b_iter = "456def".chars().peekable();
2140
2141 compare_numeric_segments(&mut a_iter, &mut b_iter);
2142
2143 assert_eq!(a_iter.collect::<String>(), "abc");
2144 assert_eq!(b_iter.collect::<String>(), "def");
2145 }
2146
2147 #[perf]
2148 fn test_natural_sort() {
2149 // Basic alphanumeric
2150 assert_eq!(natural_sort("a", "b"), Ordering::Less);
2151 assert_eq!(natural_sort("b", "a"), Ordering::Greater);
2152 assert_eq!(natural_sort("a", "a"), Ordering::Equal);
2153
2154 // Case sensitivity
2155 assert_eq!(natural_sort("a", "A"), Ordering::Less);
2156 assert_eq!(natural_sort("A", "a"), Ordering::Greater);
2157 assert_eq!(natural_sort("aA", "aa"), Ordering::Greater);
2158 assert_eq!(natural_sort("aa", "aA"), Ordering::Less);
2159
2160 // Numbers
2161 assert_eq!(natural_sort("1", "2"), Ordering::Less);
2162 assert_eq!(natural_sort("2", "10"), Ordering::Less);
2163 assert_eq!(natural_sort("02", "10"), Ordering::Less);
2164 assert_eq!(natural_sort("02", "2"), Ordering::Greater);
2165
2166 // Mixed alphanumeric
2167 assert_eq!(natural_sort("a1", "a2"), Ordering::Less);
2168 assert_eq!(natural_sort("a2", "a10"), Ordering::Less);
2169 assert_eq!(natural_sort("a02", "a2"), Ordering::Greater);
2170 assert_eq!(natural_sort("a1b", "a1c"), Ordering::Less);
2171
2172 // Multiple numeric segments
2173 assert_eq!(natural_sort("1a2", "1a10"), Ordering::Less);
2174 assert_eq!(natural_sort("1a10", "1a2"), Ordering::Greater);
2175 assert_eq!(natural_sort("2a1", "10a1"), Ordering::Less);
2176
2177 // Special characters
2178 assert_eq!(natural_sort("a-1", "a-2"), Ordering::Less);
2179 assert_eq!(natural_sort("a_1", "a_2"), Ordering::Less);
2180 assert_eq!(natural_sort("a.1", "a.2"), Ordering::Less);
2181
2182 // Unicode
2183 assert_eq!(natural_sort("文1", "文2"), Ordering::Less);
2184 assert_eq!(natural_sort("文2", "文10"), Ordering::Less);
2185 assert_eq!(natural_sort("🔤1", "🔤2"), Ordering::Less);
2186
2187 // Empty and special cases
2188 assert_eq!(natural_sort("", ""), Ordering::Equal);
2189 assert_eq!(natural_sort("", "a"), Ordering::Less);
2190 assert_eq!(natural_sort("a", ""), Ordering::Greater);
2191 assert_eq!(natural_sort(" ", " "), Ordering::Less);
2192
2193 // Mixed everything
2194 assert_eq!(natural_sort("File-1.txt", "File-2.txt"), Ordering::Less);
2195 assert_eq!(natural_sort("File-02.txt", "File-2.txt"), Ordering::Greater);
2196 assert_eq!(natural_sort("File-2.txt", "File-10.txt"), Ordering::Less);
2197 assert_eq!(natural_sort("File_A1", "File_A2"), Ordering::Less);
2198 assert_eq!(natural_sort("File_a1", "File_A1"), Ordering::Less);
2199 }
2200
2201 #[perf]
2202 fn test_compare_paths() {
2203 // Helper function for cleaner tests
2204 fn compare(a: &str, is_a_file: bool, b: &str, is_b_file: bool) -> Ordering {
2205 compare_paths((Path::new(a), is_a_file), (Path::new(b), is_b_file))
2206 }
2207
2208 // Basic path comparison
2209 assert_eq!(compare("a", true, "b", true), Ordering::Less);
2210 assert_eq!(compare("b", true, "a", true), Ordering::Greater);
2211 assert_eq!(compare("a", true, "a", true), Ordering::Equal);
2212
2213 // Files vs Directories
2214 assert_eq!(compare("a", true, "a", false), Ordering::Greater);
2215 assert_eq!(compare("a", false, "a", true), Ordering::Less);
2216 assert_eq!(compare("b", false, "a", true), Ordering::Less);
2217
2218 // Extensions
2219 assert_eq!(compare("a.txt", true, "a.md", true), Ordering::Greater);
2220 assert_eq!(compare("a.md", true, "a.txt", true), Ordering::Less);
2221 assert_eq!(compare("a", true, "a.txt", true), Ordering::Less);
2222
2223 // Nested paths
2224 assert_eq!(compare("dir/a", true, "dir/b", true), Ordering::Less);
2225 assert_eq!(compare("dir1/a", true, "dir2/a", true), Ordering::Less);
2226 assert_eq!(compare("dir/sub/a", true, "dir/a", true), Ordering::Less);
2227
2228 // Case sensitivity in paths
2229 assert_eq!(
2230 compare("Dir/file", true, "dir/file", true),
2231 Ordering::Greater
2232 );
2233 assert_eq!(
2234 compare("dir/File", true, "dir/file", true),
2235 Ordering::Greater
2236 );
2237 assert_eq!(compare("dir/file", true, "Dir/File", true), Ordering::Less);
2238
2239 // Hidden files and special names
2240 assert_eq!(compare(".hidden", true, "visible", true), Ordering::Less);
2241 assert_eq!(compare("_special", true, "normal", true), Ordering::Less);
2242 assert_eq!(compare(".config", false, ".data", false), Ordering::Less);
2243
2244 // Mixed numeric paths
2245 assert_eq!(
2246 compare("dir1/file", true, "dir2/file", true),
2247 Ordering::Less
2248 );
2249 assert_eq!(
2250 compare("dir2/file", true, "dir10/file", true),
2251 Ordering::Less
2252 );
2253 assert_eq!(
2254 compare("dir02/file", true, "dir2/file", true),
2255 Ordering::Greater
2256 );
2257
2258 // Root paths
2259 assert_eq!(compare("/a", true, "/b", true), Ordering::Less);
2260 assert_eq!(compare("/", false, "/a", true), Ordering::Less);
2261
2262 // Complex real-world examples
2263 assert_eq!(
2264 compare("project/src/main.rs", true, "project/src/lib.rs", true),
2265 Ordering::Greater
2266 );
2267 assert_eq!(
2268 compare(
2269 "project/tests/test_1.rs",
2270 true,
2271 "project/tests/test_2.rs",
2272 true
2273 ),
2274 Ordering::Less
2275 );
2276 assert_eq!(
2277 compare(
2278 "project/v1.0.0/README.md",
2279 true,
2280 "project/v1.10.0/README.md",
2281 true
2282 ),
2283 Ordering::Less
2284 );
2285 }
2286
2287 #[perf]
2288 fn test_natural_sort_case_sensitivity() {
2289 std::thread::sleep(std::time::Duration::from_millis(100));
2290 // Same letter different case - lowercase should come first
2291 assert_eq!(natural_sort("a", "A"), Ordering::Less);
2292 assert_eq!(natural_sort("A", "a"), Ordering::Greater);
2293 assert_eq!(natural_sort("a", "a"), Ordering::Equal);
2294 assert_eq!(natural_sort("A", "A"), Ordering::Equal);
2295
2296 // Mixed case strings
2297 assert_eq!(natural_sort("aaa", "AAA"), Ordering::Less);
2298 assert_eq!(natural_sort("AAA", "aaa"), Ordering::Greater);
2299 assert_eq!(natural_sort("aAa", "AaA"), Ordering::Less);
2300
2301 // Different letters
2302 assert_eq!(natural_sort("a", "b"), Ordering::Less);
2303 assert_eq!(natural_sort("A", "b"), Ordering::Less);
2304 assert_eq!(natural_sort("a", "B"), Ordering::Less);
2305 }
2306
2307 #[perf]
2308 fn test_natural_sort_with_numbers() {
2309 // Basic number ordering
2310 assert_eq!(natural_sort("file1", "file2"), Ordering::Less);
2311 assert_eq!(natural_sort("file2", "file10"), Ordering::Less);
2312 assert_eq!(natural_sort("file10", "file2"), Ordering::Greater);
2313
2314 // Numbers in different positions
2315 assert_eq!(natural_sort("1file", "2file"), Ordering::Less);
2316 assert_eq!(natural_sort("file1text", "file2text"), Ordering::Less);
2317 assert_eq!(natural_sort("text1file", "text2file"), Ordering::Less);
2318
2319 // Multiple numbers in string
2320 assert_eq!(natural_sort("file1-2", "file1-10"), Ordering::Less);
2321 assert_eq!(natural_sort("2-1file", "10-1file"), Ordering::Less);
2322
2323 // Leading zeros
2324 assert_eq!(natural_sort("file002", "file2"), Ordering::Greater);
2325 assert_eq!(natural_sort("file002", "file10"), Ordering::Less);
2326
2327 // Very large numbers
2328 assert_eq!(
2329 natural_sort("file999999999999999999999", "file999999999999999999998"),
2330 Ordering::Greater
2331 );
2332
2333 // u128 edge cases
2334
2335 // Numbers near u128::MAX (340,282,366,920,938,463,463,374,607,431,768,211,455)
2336 assert_eq!(
2337 natural_sort(
2338 "file340282366920938463463374607431768211454",
2339 "file340282366920938463463374607431768211455"
2340 ),
2341 Ordering::Less
2342 );
2343
2344 // Equal length numbers that overflow u128
2345 assert_eq!(
2346 natural_sort(
2347 "file340282366920938463463374607431768211456",
2348 "file340282366920938463463374607431768211455"
2349 ),
2350 Ordering::Greater
2351 );
2352
2353 // Different length numbers that overflow u128
2354 assert_eq!(
2355 natural_sort(
2356 "file3402823669209384634633746074317682114560",
2357 "file340282366920938463463374607431768211455"
2358 ),
2359 Ordering::Greater
2360 );
2361
2362 // Leading zeros with numbers near u128::MAX
2363 assert_eq!(
2364 natural_sort(
2365 "file0340282366920938463463374607431768211455",
2366 "file340282366920938463463374607431768211455"
2367 ),
2368 Ordering::Greater
2369 );
2370
2371 // Very large numbers with different lengths (both overflow u128)
2372 assert_eq!(
2373 natural_sort(
2374 "file999999999999999999999999999999999999999999999999",
2375 "file9999999999999999999999999999999999999999999999999"
2376 ),
2377 Ordering::Less
2378 );
2379 }
2380
2381 #[perf]
2382 fn test_natural_sort_case_sensitive() {
2383 // Numerically smaller values come first.
2384 assert_eq!(natural_sort("File1", "file2"), Ordering::Less);
2385 assert_eq!(natural_sort("file1", "File2"), Ordering::Less);
2386
2387 // Numerically equal values: the case-insensitive comparison decides first.
2388 // Case-sensitive comparison only occurs when both are equal case-insensitively.
2389 assert_eq!(natural_sort("Dir1", "dir01"), Ordering::Less);
2390 assert_eq!(natural_sort("dir2", "Dir02"), Ordering::Less);
2391 assert_eq!(natural_sort("dir2", "dir02"), Ordering::Less);
2392
2393 // Numerically equal and case-insensitively equal:
2394 // the lexicographically smaller (case-sensitive) one wins.
2395 assert_eq!(natural_sort("dir1", "Dir1"), Ordering::Less);
2396 assert_eq!(natural_sort("dir02", "Dir02"), Ordering::Less);
2397 assert_eq!(natural_sort("dir10", "Dir10"), Ordering::Less);
2398 }
2399
2400 #[perf]
2401 fn test_natural_sort_edge_cases() {
2402 // Empty strings
2403 assert_eq!(natural_sort("", ""), Ordering::Equal);
2404 assert_eq!(natural_sort("", "a"), Ordering::Less);
2405 assert_eq!(natural_sort("a", ""), Ordering::Greater);
2406
2407 // Special characters
2408 assert_eq!(natural_sort("file-1", "file_1"), Ordering::Less);
2409 assert_eq!(natural_sort("file.1", "file_1"), Ordering::Less);
2410 assert_eq!(natural_sort("file 1", "file_1"), Ordering::Less);
2411
2412 // Unicode characters
2413 // 9312 vs 9313
2414 assert_eq!(natural_sort("file①", "file②"), Ordering::Less);
2415 // 9321 vs 9313
2416 assert_eq!(natural_sort("file⑩", "file②"), Ordering::Greater);
2417 // 28450 vs 23383
2418 assert_eq!(natural_sort("file漢", "file字"), Ordering::Greater);
2419
2420 // Mixed alphanumeric with special chars
2421 assert_eq!(natural_sort("file-1a", "file-1b"), Ordering::Less);
2422 assert_eq!(natural_sort("file-1.2", "file-1.10"), Ordering::Less);
2423 assert_eq!(natural_sort("file-1.10", "file-1.2"), Ordering::Greater);
2424 }
2425
2426 #[test]
2427 fn test_multiple_extensions() {
2428 // No extensions
2429 let path = Path::new("/a/b/c/file_name");
2430 assert_eq!(path.multiple_extensions(), None);
2431
2432 // Only one extension
2433 let path = Path::new("/a/b/c/file_name.tsx");
2434 assert_eq!(path.multiple_extensions(), None);
2435
2436 // Stories sample extension
2437 let path = Path::new("/a/b/c/file_name.stories.tsx");
2438 assert_eq!(path.multiple_extensions(), Some("stories.tsx".to_string()));
2439
2440 // Longer sample extension
2441 let path = Path::new("/a/b/c/long.app.tar.gz");
2442 assert_eq!(path.multiple_extensions(), Some("app.tar.gz".to_string()));
2443 }
2444
2445 #[test]
2446 fn test_strip_path_suffix() {
2447 let base = Path::new("/a/b/c/file_name");
2448 let suffix = Path::new("file_name");
2449 assert_eq!(strip_path_suffix(base, suffix), Some(Path::new("/a/b/c")));
2450
2451 let base = Path::new("/a/b/c/file_name.tsx");
2452 let suffix = Path::new("file_name.tsx");
2453 assert_eq!(strip_path_suffix(base, suffix), Some(Path::new("/a/b/c")));
2454
2455 let base = Path::new("/a/b/c/file_name.stories.tsx");
2456 let suffix = Path::new("c/file_name.stories.tsx");
2457 assert_eq!(strip_path_suffix(base, suffix), Some(Path::new("/a/b")));
2458
2459 let base = Path::new("/a/b/c/long.app.tar.gz");
2460 let suffix = Path::new("b/c/long.app.tar.gz");
2461 assert_eq!(strip_path_suffix(base, suffix), Some(Path::new("/a")));
2462
2463 let base = Path::new("/a/b/c/long.app.tar.gz");
2464 let suffix = Path::new("/a/b/c/long.app.tar.gz");
2465 assert_eq!(strip_path_suffix(base, suffix), Some(Path::new("")));
2466
2467 let base = Path::new("/a/b/c/long.app.tar.gz");
2468 let suffix = Path::new("/a/b/c/no_match.app.tar.gz");
2469 assert_eq!(strip_path_suffix(base, suffix), None);
2470
2471 let base = Path::new("/a/b/c/long.app.tar.gz");
2472 let suffix = Path::new("app.tar.gz");
2473 assert_eq!(strip_path_suffix(base, suffix), None);
2474 }
2475
2476 #[cfg(target_os = "windows")]
2477 #[test]
2478 fn test_wsl_path() {
2479 use super::WslPath;
2480 let path = "/a/b/c";
2481 assert_eq!(WslPath::from_path(&path), None);
2482
2483 let path = r"\\wsl.localhost";
2484 assert_eq!(WslPath::from_path(&path), None);
2485
2486 let path = r"\\wsl.localhost\Distro";
2487 assert_eq!(
2488 WslPath::from_path(&path),
2489 Some(WslPath {
2490 distro: "Distro".to_owned(),
2491 path: "/".into(),
2492 })
2493 );
2494
2495 let path = r"\\wsl.localhost\Distro\blue";
2496 assert_eq!(
2497 WslPath::from_path(&path),
2498 Some(WslPath {
2499 distro: "Distro".to_owned(),
2500 path: "/blue".into()
2501 })
2502 );
2503
2504 let path = r"\\wsl$\archlinux\tomato\.\paprika\..\aubergine.txt";
2505 assert_eq!(
2506 WslPath::from_path(&path),
2507 Some(WslPath {
2508 distro: "archlinux".to_owned(),
2509 path: "/tomato/paprika/../aubergine.txt".into()
2510 })
2511 );
2512
2513 let path = r"\\windows.localhost\Distro\foo";
2514 assert_eq!(WslPath::from_path(&path), None);
2515 }
2516}